[rbak-nsp] BGP next hop as route server IP

[Jeff Tantsura]

Hi,

Is this issue only reproducible with 6.2.1.4?
What does "sometimes" mean, is there some condition to provoke this?
Have you tried to debug?
Are you sure the RS "sometimes" doesn't rewrite NH to itself?

How does your config look like?

Cheers,
Jeff

On Wed, Oct 20, 2010 at 2:20 AM, Marcin Kuczera <marcin at leon.pl> wrote:

> hello,
>
> we are having recently problems with connection to RouteServers.
> Well, RS'es are configured to add their AS as first in path, so connection
> is possible.
>
> However, sometimes it happens, that all next hops are set as RouteServer IP
> addresses (and so traffic is dropped by RS).
>
> It happens more often, when we have iBGP between 2 SE100, and both SE100
> are connected paralelly to the same route-servers (via the same VLAN, but
> different IP addresses from the same /23 network).
>
> If there are some problems with RSes, or one of sessions goes down - SE100
> with 6.2.1.4 locks next hop as RS address.
> SoftClear, clear, whatever - doesn't help.
> Restart of bgp process, rib process - also, no result.
>
> The only solution is to remove IP address from IX interface (PLIX), clear
> BGP sessions (with 2 RSes, on both the same problem same time).
> Then add this IP back to PLIX interface and commit.
>
> (and in meantime, SEOS sends some ARP requests from other VLANs to PLIX
> vlan !?!?)
>
>
>
> Is it a bug or there is some condition at which it is possible to replace
> proper next-hop with RS IP ?
>
> (RSes advertise proper next-hop addresses)
>
> Regards,
> Marcin
> _______________________________________________
> redback-nsp mailing list
> redback-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/redback-nsp
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://puck.nether.net/pipermail/redback-nsp/attachments/20101021/1db98e56/attachment-0001.html>